Quantitative Evaluation of Encrustations in Double-J Ureteral Stents With Micro-computed Tomography and Semantic Segmentation
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Quantitative Evaluation of Encrustations in Double-J Ureteral Stents With Micro-computed Tomography and Semantic Segmentation Shaokai Zheng ( shaokai.zheng@outlook.com ) University of Bern Pedro Amado University of Bern Bernhard Kiss University Hospital of Bern Fabian Stangl University Hospital of Bern Andreas Haeberlin University Hospital of Bern Daniel Sidler University Hospital of Bern Dominik Obrist University of Bern Fiona Burkhard University Hospital of Bern Francesco Clavica University of Bern Research Article Keywords: Double-J ureteral stents, micro-computed tomography, semantic segmentation, Quantitative evaluation, encrustations Posted Date: August 18th, 2021 DOI: https://doi.org/10.21203/rs.3.rs-763507/v1 License: This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 1/13
Abstract Accurate evaluations of stent encrustation patterns, such as volume distribution, from different patient groups are valuable for clinical management and the development of better stents. This study compared stent encrustation patterns from stone and kidney transplant patients. Twenty-three double-J ureteral stents were collected at a single center from patients with stone disease or underwent kidney transplantation. Encrustations on stent samples were quantified by means of micro‑computed tomography and semantic segmentation using Convolutional Neural Network models. Luminal encrustation volume per stent unit was derived to represent encrustation level, which did not differ between patient groups in the first six weeks. However, stone patients showed higher encrustation levels over prolonged indwelling times (p = 0.036). Along the stent shaft body, the stone group showed higher encrustation levels near the ureteropelvic junction compared to the ureterovesical junction (p = 0.013), whereas the transplant group showed no such difference. Possible explanations were discussed regarding vesicoureteral refluxes. In both patient groups, stent pigtails were more susceptible to encrustations, and no difference between renal and bladder pigtail was identified. Our results suggest that excessively long stents with superfluous pigtails should be avoided. Introduction Double-J ureteral stents are commonly used for pain management in acute obstruction prior to endoscopic stone treatment, or to stent the ureterovesical anastomosis after kidney transplantation to avoid obstruction due to swelling in the early postoperative phase. In spite of various material upgrades and design modifications, encrustation remains a major problem causing stent associated complications with a significant impact on patients’ quality of life 1,2. Encrusted stents can become obstructed and lead to infections including pyelonephritis. Severely encrusted stents can no longer be retrieved cystoscopically and require more invasive approaches to be removed. As a result, indwelling stents have to be replaced in regular intervals. For stone patients, indwelling times longer than six weeks have been associated with significantly higher encrustation rates2–4. For transplant patients, indwelling times from two to six weeks after transplantation have all been advised based on urinary tract infection (UTI) rates and stent-related complications such as pain, haematuria, encrustation and migration5–8. In this study, we are interested in clarifying the encrustation level over indwelling time in both stone and transplant patient groups, with an emphasis on the localization of the encrustations in each group. Therefore, evaluation of the encrustation volume is crucial. Earlier approaches were mainly qualitative, relying on visual examination of Scanning Electron Microscopy9 or kidney-ureter-bladder (KUB) radiography images10,11. Extracting quantitative information from KUB images had been attempted by measuring the projected area of encrustations12. Unfortunately, the inherent uncertainty is not negligible as encrustations are three-dimensional. Another established approach proposed to measure the level of encrustation by weighing the stent sample before and after oxidative acid treatment to dissolve the encrustations13. The spatial distribution of encrustations, however, is destroyed during the process. Page 2/13
Practically, encrustations on external stent surfaces can be significantly affected during removal and consequently introduce significant uncertainties to quantitative results. It is therefore desirable to isolate the luminal encrustations (in the stent lumen), which are less affected by stent removal and critical for luminal blockage of indwelling stents. In a recently published work, micro-computed tomography (µ-CT) was applied to analyse stent encrustation volumes to assess the anti-encrustation efficacy of two commercially available stents14. The authors performed morphological segmentation on the µ-CT images, and managed to distinguish between luminal and external encrustations. Based on their results, more than 90% of the stent had luminal encrustations, which also appeared more in the shaft body (the straight part of the stent) than in the renal and bladder pigtails, respectively. Nonetheless, one limitation associated with the method was that the volume of individual stent remained unknown, so the encrustation volume might be biased according to the actual stent sizes. Moreover, treating the entire stent shaft body as a unified section inherently ignores any heterogeneity of the distribution of encrustation volume along the shaft, which might overlook some key characteristics along the shaft body. In this study, we compared patterns of luminal stent encrustations in stone and transplant patients. A semantic segmentation approach was first evaluated and implemented by means of Convolutional Neural Network (CNN) to simultaneously quantify the volume of luminal encrustations and the stent. Luminal encrustation volume per stent unit was derived, representing the normalized encrustation level, where the inter-subject changes of stent volume were accounted for. Further, luminal encrustation levels along the stent were assessed to evaluate the heterogeneity along the stent, and comparisons were made between stone and transplant patients. Results Study population. A total of 23 stents were collected from 23 ureter units, as summarized in Table 1. Stent indwelling times were 11–90 days for stone patients and 22–42 days for kidney transplant patients, with no significant difference between the groups (p > 0.05). All retrieved stents (KUARTZ GUARDIAN Black Loop, PURE Medical Device SA, Geneva, CH) were made of thermosensitive polyurethane with hydrophilic coatings. Page 3/13
Table 1 Characteristics of patients and stents. Stone patients Transplant patients Patients, n 13 9 Female, n (%) 6 (46%) 5 (56%) Median age, yr (IQR) 50.5 (47–69) 56 (38-73.5) Presence of UTI, n (%) 3/13 (23%) 3/9 (33%) Stone location, n (%) Nephrolithiasis 9/13 (69.2%) NA Ureterolithiasis 3/13 (23.1%) NA Nephro- & Ureterolithiasis 1/13 (7.7%) NA Collected stents, n 14* 9 Stent size, Fr/cm 6/26 6/10 Median indwelling time, d (IQR) 32.5 (23–42) 28 (25.5–33.3) IQR = interquartile range; UTI = urinary tract infection; yr = year; d = day. * From 14 ureter units. Encrustations near side holes. Segmented images from stents with minimal, median (or nearest), and maximal indwelling times are shown in Fig. 1, respectively, for each patient group. The amount of luminal encrustations (orange) in the stone group seems to increase over time, which is not observed in the transplant group. The increasing encrustation in the stone group is more apparent on the renal pigtail and proximal straight part of the stent (see Methods section). Aggregates of encrustations were mostly found near the side holes (SHs), whose locations along the pigtails are marked (arrows) in Fig. 1. The example with indwelling time of 90 days showed complete blockage in the proximal lumen, as evidenced by the cross-sectional inset in Fig. 1. Encrustation over time. The total encrustation volume ratios (TEVRs, see Methods section) from stone patients were significantly different with indwelling times < 42 d (median: 0.45, IQR: 0.10–0.80) and ≥ 42 d (median: 8.4, IQR: 4.0–21, p = 0.036). Further comparisons of the encrustation volume ratios (EVRs, see Methods section) in each section (Fig. 2a) revealed that encrustations increased significantly in the renal pigtail (p = 0.004) and the distal straight part (p = 0.014). The EVR in the proximal straight part also showed considerable increase over six weeks but did not reach statistical significance (p = 0.054). In the transplant group no significant difference was found between short indwelling time (< 28 d) and long indwelling time (≥ 28 d) groups (see Supplementary Material S2), while the median EVRs were higher in Page 4/13
both pigtails (Fig. 2b). The comparison was also made between stents from stone and transplant groups with indwelling times < 42 d (Fig. 2c) and no differences were found in EVR (p > 0.05 in all sections) or TEVR (p = 0.3). Most encrusted stent region. Subsequent comparisons of encrustation risk levels (ERLs, see Methods section) were made between stone and transplant groups for each stent section (Fig. 2d). In both patient groups, the ERLs were higher in the pigtails with not significantly difference between renal and bladder. Interestingly, in stone patients, the ERL of the proximal straight part was higher than that of the distal part (p = 0.009). In contrast, no significant difference was found between the straight parts in transplant patients. The ERLs of the pigtails were significantly higher than in the proximal (p = 0.028) and distal (p = 0.031) straight parts, respectively. Comparison of ERLs between stone and transplant groups revealed a significant difference in the proximal straight part, with the higher ERL in the stone patients (p = 0.013). Further data can be found in the Supplementary Material S2. Discussion The CNN model15 based on deep-learning allowed us to evaluate luminal encrustation and stent volumes simultaneously with unparalleled accuracy. Once trained, the model can be applied to subsequent data sets without further tuning, and therefore reduces random error or bias imposed during the quantitative analyses. Moreover, to the best authors’ knowledge, this study is the first to measure both the encrustation volume and the stent volume. As such, the encrustation volume per stent unit was calculated, which is more representative than directly comparing the encrustation volumes, as was done in several previous studies13,14. One observation in our study was that SHs were often the anchoring sites of encrustations even for stents with short indwelling times regardless of the patient group (Fig. 1). The initial deposits of encrustations near SHs might be explained by recent in vitro experiments such that SHs facilitate local urine flow stasis and promote particle accumulation in the neighbouring regions16–18. These initial deposits exacerbate the local encrustation process, causing severe stone burden near SHs11,19. The stony encrustation could compromise the stent tensile strength at the SHs, which deteriorate into fractures11,19. Another observation was that that encrustation levels did not differ between patient groups in the first six weeks, but stone patients had a higher tendency to build up encrustations over prolonged indwelling time than transplant patients. We also observed the highest ERLs in the pigtails in both patient groups, suggesting higher risks of excessive encrustations per stent unit. Both pigtails exhibited high median in ERL and did not differ from each other, so the risk seemed equivalent in the renal and bladder ends. Since our results were based on encrustation volume per stent unit, the use of excessively long stents with superfluous pigtail materials should be avoided in both patient groups. In practice, proximal stone burden has been associated with multiple surgical complications 20, and accurate determination of the stent section most susceptible to encrustations has been an active topic of Page 5/13
discussion. Previous studies on stent encrustation patterns mainly focused on stone patients and no consensus was reached. While some suggested the renal pigtail as the most susceptible section followed by the bladder pigtail4,11,13, others reported that the two pigtails or the distal part of the stent were equally susceptible9, or that the distal part of stent12 was most encrusted. The recent study using µ-CT14 suggested the stent’s shaft body (the entire straight part) as the most susceptible region. However, no previous efforts attempted to derive encrustation volume per stent unit as an endpoint, which should be more meaningful than direct comparison of encrustation volumes. For one thing, the disputed conclusions can be attributed to the lack of quantitative tools to accurately measure the encrustation volumes. For another, the fact that there were no significant differences between certain comparisons, as demonstrated in Fig. 2, should also be acknowledged. Other than stone patients, quantitative evaluation of stent encrustation patterns in transplant patients are still lacking to the best of the authors’ knowledge. Following kidney transplantation, stents are usually placed to prevent strictures or urine leakage. These prophylactic stents are usually much shorter than those in stone patients since the allograft ureter is kept short to ensure a good blood supply. Consequently, vesicoureteral reflux (VUR) in transplant patients often reaches up to the renal pelvis21. This retrograde flow might create a flushing effect, periodically removing the deposits on the luminal surface of the stent, whereas in stone patients the VUR would not reach up to renal pelvis as easily since the stents are often longer. This might explain the higher encrustation level in the proximal straight part in stone patients. Moreover, stent implantation significantly impedes the peristalsis of the ureter22, and thus the urine transport from kidney to bladder is largely passive. The longer tube in stone patients creates larger resistance to the fluid, and local stasis of urine can be formed near the UPJ as the tract narrows, whereas the shorter tube in transplant patients could better facilitate urine flows through the stent, either from kidney to bladder or in presence of VUR. As such, both forward and retrograde urine flows influences the dynamics, co-creating the different encrustation patterns presented above. In terms of limitation, the current study consists of stents collected from a single center with limited sample size, preventing further subgroup analysis. Moreover, the mechanism of urine flow dynamics in stented ureters remains unclear, which plays a pivotal role in stent encrustation23. Further studies in this regard could help elucidate the process of encrustations in stented native or allograft ureters. Conclusion The semantic segmentation approach delivered accurate and intuitive results and is worth further applications to image analyses in urology. Our results highlighted the similarities and differences in encrustation patterns between stone and renal transplant groups, and possible explanations were discussed. Further investigations in both engineering and clinical disciplines are required to fully understand the dynamics of encrustations in order to develop the “perfect stent”. Page 6/13
Materials And Methods Study guidelines and ethics. The retrospective study was reported in accordance with guidelines from the STROBE statement24. The stone and transplant cohorts consisted of double-J stents removed endoscopically from patients underwent stone treatment and kidney transplantation, respectively, between January and December 2020 at the Department of Urology, Bern University Hospital. Only stents manufactured by PURE Medical Device SA were included in the study. Size of stents were limited to 6Fr/26cm and 6Fr/10cm in the stone and transplant cohorts, respectively. The age, gender, stent indwelling time from the transplant cohort were matched with that of the stone cohort (p > 0.05). Exclusion criteria included: (1) stents placed for external obstructions such as pregnancy and urothelial carcinoma; and (2) stents with unknown indwelling times. Since all data were collected as by-products of regular urological treatment and anonymized, this study did not fall within the scope of the Swiss Federal Act on Research involving Human Beings (Human Research Act, Art. 2, The Federal Assembly of the Swiss Confederation), and therefore did not need approval of the local ethics committee. General consent (v5.0 February 2016, Bern University Hospital, Directorate of Teaching and Research, Insel Gruppe AG) was obtained from all patients. Informed consent was waived under the Human Research Act Art. 34. All methods were carried out in accordance with relevant guidelines and regulations. Material preparation. Each collected stent was dried in an oven at 60°C for three hours to remove residual urine To quantify the encrustation volume in different sections of the stent, each stent was divided into four sections, that is, the renal pigtail, the proximal straight part (near the UPJ), the distal straight part (near the UVJ), and the bladder pigtail (see Fig. 3a). This was done under the assumption that the two junctions are critical regions for the development of encrustations as they are the entrance and exist of the ureter, where urine flows are regulated by the physiologically narrowing tract. Imaging and segmentation. Subsequently, each section was scanned using a µ-CT scanner (SCANCO Medical AG, Bruettisellen, CH) to acquire a three-dimensional image, which was segmented using a CNN model known as U-Net15, available in the Dragonfly software (v2020.1, Object Research Systems Inc. Quebec, CA, http://www.theobjects.com/dragonfly). The segmentation result allowed evaluation of the luminal encrustations including the luminal space of the SHs without losing their spatial distribution (see Fig. 3b-e), thus offering a more reliable representation of the encrustations14. Full technical details on µ- CT, discussion on accuracy, and examples can be found in the Supplementary Material S1. Study outcomes and definitions. By defining the encrustation volume ratio (EVR) as the encrustation volume normalized by the corresponding stent volume, which gives the encrustation volume per stent unit, the bias introduced by the different volumes of individual stent samples was eliminated. The total encrustation volume ratio (TEVR) was defined by summing EVR over the four stent sections. To study the relative level of encrustations at different sections, we defined the encrustation risk level (ERL) by dividing EVR over TEVR. The stent section with highest ERL would be most likely to attract encrustations. Page 7/13
To compare the encrustation level over time, stents from stone patients were divided into 2 groups: group 1 with indwelling time < 42 d, and group 2 with indwelling time ≥ 42 d. This discrimination was based on the fact that an indwelling time over six weeks is commonly associated with significantly higher encrustation levels2–4. Since optimal removal time in transplant patients has been reported to range between two and six weeks and no consensus has been reached, subgroups with indwelling times < 28 d or ≥ 28 d were chosen for comparison. Statistical analysis. For statistical comparisons, median values with interquartile range (IQR) were used, and the two-sided Mann-Whitney U-test was used for significance tests. P-values from multiple comparisons were corrected using the Bonferroni-Holm method, and p < 0.05 was considered significant in this study. Declarations Acknowledgement This publication is based upon work from COST Action "European Network of Multidisciplinary Research to Improve the Urinary Stents - CA16217", supported by COST (European Cooperation in Science and Technology). The authors also acknowledge the financial support from Swiss National Science Foundation (SNSF grant number IZCOZ0_182966). S.Z. acknowledges Mr. Michael Indermaur from the Musculoskeletal Biomechanics group at ARTORG for his support in operating the Micro-CT scanner, and Mr. Emile Talon for his contribution in the preliminary study. Author contributions Conception and design: F.B., F.C., D.O., S.Z.; Acquisition of data: P.A., F.B., B.K., F.S., S.Z.; Analysis and interpretation of data: P.A., S.Z.; Figure preparation: P.A., S.Z; Drafting of the manuscript: S.Z.; Statistical analysis: A.H., S.Z.; All authors provided critical revision of the manuscript. Competing interests The authors declare no competing interests. Additional Information Supplementary materials are available online. Further requests on technical details should be addressed to S.Z. References 1. Bibby, L. M. & Wiseman, O. J. Double JJ Ureteral Stenting: Encrustation and Tolerability. European Urology Focus 7, 7-8, doi:10.1016/j.euf.2020.08.014 (2020). Page 8/13
2. Tomer, N., Garden, E., Small, A. & Palese, M. Ureteral Stent Encrustation: Epidemiology, Pathophysiology, Management and Current Technology. Journal of Urology 205, 68-77, doi:10.1097/JU.0000000000001343 (2021). 3. El-Faqih, S. R. et al. Polyurethane Internal Ureteral Stents in Treatment of Stone Patients: Morbidity Related to Indwelling Times. The Journal of Urology 146, 1487-1491, doi:https://doi.org/10.1016/S0022-5347(17)38146-6 (1991). 4. Kawahara, T., Ito, H., Terao, H., Yoshida, M. & Matsuzaki, J. Ureteral Stent Encrustation, Incrustation, and Coloring: Morbidity Related to Indwelling Times. Journal of Endourology 26, 178-182, doi:10.1089/end.2011.0385 (2011). 5. Tavakoli, A. et al. Impact of Stents on Urological Complications and Health Care Expenditure in Renal Transplant Recipients: Results of a Prospective, Randomized Clinical Trial. Journal of Urology 177, 2260-2264, doi:10.1016/j.juro.2007.01.152 (2007). 6. Thompson, E. R., Hosgood, S. A., Nicholson, M. L. & Wilson, C. H. Early versus late ureteric stent removal after kidney transplantation. Cochrane Database of Systematic Reviews, doi:10.1002/14651858.CD011455.pub2 (2018). 7. Visser, I. J. et al. Timing of Ureteric Stent Removal and Occurrence of Urological Complications after Kidney Transplantation: A Systematic Review and Meta-Analysis. Journal of Clinical Medicine 8, doi:10.3390/jcm8050689 (2019). 8. Wilson, C. H., Rix, D. A. & Manas, D. M. Routine intraoperative ureteric stenting for kidney transplant recipients. Cochrane Database of Systematic Reviews, doi:10.1002/14651858.CD004925.pub3 (2013). 9. Arkusz, K., Pasik, K., Halinski, A. & Halinski, A. Surface analysis of ureteral stent before and after implantation in the bodies of child patients. Urolithiasis, doi:10.1007/s00240-020-01211-9 (2020). 10. Acosta-Miranda, A. M., Milner, J. & Turk, T. M. T. The FECal Double-J: A Simplified Approach in the Management of Encrusted and Retained Ureteral Stents. Journal of Endourology 23, 409-415, doi:10.1089/end.2008.0214 (2009). 11. Singh, I. et al. Severely encrusted polyurethane ureteral stents: management and analysis of potential risk factors. Urology 58, 526-531, doi:10.1016/S0090-4295(01)01317-6 (2001). 12. Rana, A. M. & Sabooh, A. Management Strategies and Results for Severely Encrusted Retained Ureteral Stents. Journal of Endourology 21, 628-632, doi:10.1089/end.2006.0250 (2007). 13. Sighinolfi, M. C. et al. Chemical and Mineralogical Analysis of Ureteral Stent Encrustation and Associated Risk Factors. Urology 86, 703-706, doi: 10.1016/j.urology.2015.05.015 (2015). 14. Yoshida, T. et al. A randomized clinical trial evaluating the short-term results of ureteral stent encrustation in urolithiasis patients undergoing ureteroscopy: micro-computed tomography evaluation. Scientific Reports 11, 10337, doi:10.1038/s41598-021-89808-x (2021). 15. Ronneberger, O., Fischer, P. & Brox, T. in Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. (eds Nassir Navab, Joachim Hornegger, William M. Wells, & Alejandro F. Frangi) 234-241 (Springer International Publishing). Page 9/13
16. Mosayyebi, A. et al. Particle Accumulation in Ureteral Stents Is Governed by Fluid Dynamics: In Vitro Study Using a “Stent-on-Chip” Model. Journal of Endourology 32, 639-646 (2018). 17. Clavica, F. et al. Investigating the flow dynamics in the obstructed and stented ureter by means of a biomimetic artificial model. PloS one 9, e87433 (2014). 18. Carugo, D., Zhang, X., Drake, J. M. & Clavica, F. in Proceedings of the 18th International Conference on Miniaturized Systems for Chemistry and Life Sciences, MicroTAS. 19. Zisman, A., Siegel, Y. I., Siegmann, A. & Lindner, A. Spontaneous Ureteral Stent Fragmentation. Journal of Urology 153, 718-721, doi:10.1016/S0022-5347(01)67697-3 (1995). 20. Weedin, J. W., Coburn, M. & Link, R. E. The Impact of Proximal Stone Burden on the Management of Encrusted and Retained Ureteral Stents. The Journal of Urology 185, 542-547, doi:https://doi.org/10.1016/j.juro.2010.09.085 (2011). 21. Ness, D. & Olsburgh, J. UTI in kidney transplant. World Journal of Urology 38, 81-88, doi:10.1007/s00345-019-02742-6 (2020). 22. Mosli Hisham, A., Farsi Hasan, M. A., Al-Zimaity Mohammed, F., Saleh Tarik, R. & Al-Zamzami Mokhtar, M. Vesicoureteral Reflux in Patients with Double Pigtail Stents. Journal of Urology 146, 966- 969, doi:10.1016/S0022-5347(17)37976-4 (1991). 23. Zheng, S. et al. Fluid mechanical modeling of the upper urinary tract. WIREs Mechanisms of Disease n/a, e01523, doi:10.1002/wsbm.1523 (2021). 24. Vandenbroucke, J. P. et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): Explanation and Elaboration. PLOS Medicine 4, e297, doi:10.1371/journal.pmed.0040297 (2007). Figures Page 10/13
Figure 1 Examples of luminal encrustations (orange) in the renal pigtail, proximal straight part, distal straight part, and bladder pigtail of the stents from stone and transplant patient groups. Stents are rendered semi- transparent for better visualization. Side hole locations on the pigtails are marked by arrows, where yellow arrows indicate clear evidence of encrustations and white arrows suggest otherwise. Cross sectional insets are given for visually significant encrustations to show the blockage level in the stent lumen. Page 11/13
Figure 2 Encrustation volume ratios (EVR) compared with respect to indwelling time for (a) stone patients and (b) transplant patients. (c) EVR compared between patient groups with stent indwelling time < 42 days. (d) Encrustation risk levels (ERL) from stone and transplant patients for each stent section. The median (bar) and interquartile range (error bars) are shown. Raw data from the same stent are coded by color (filled Page 12/13
circles). P-values corrected by the Bonferroni-Holm method is indicated by the asteroid (*). Further data can be found in Supplementary Material S2. Figure 3 (a) Illustration of the four divided sections from a ureteral stent. Three dimensional µ-CT images from the pigtail and straight part before (b, c) and after (d, e) segmentations, respectively. Luminal encrustations from the segmentation results are shown in orange, where stents are rendered semi-transparent for better visualization. Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. SupplementaryMaterialS1.docx SupplementaryMaterialS2.docx Page 13/13
You can also read